Inference of signaling pathway activity in single cell with SaaSc

Abstract

Inference of pathway activity is crucial for decoding cellular heterogeneity from single-cell data. However, estimating pathway activity in individual cells presents significant challenges due to two key factors: (1) the high noise and sparsity inherent in single-cell data, and (2) the complex dependencies between pathways. To address these challenges, we have developed SaaSc, a tool for estimating pathway activity at single-cell resolution. SaaSc employs multiple correspondence analysis (MCA) to decompose single-cell data and reconstruct gene expression matrices by selecting the top dimensions, thereby increasing the signal-to-noise ratio. Additionally, we utilized ridge regression, which enhances the performance of linear regression models, to directly infer pathway activity.

Install

install.packages("devtools")
devtools::install_github("yoyoong/SaaSc", ref = "main")